OCCAM: Open-set Causal Concept explAnation and Ontology induction for black-box vision Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910232602476544 |
|---|---|
| author | Russo, Chiara Maria Carnemolla, Simone Palazzo, Simone Giordano, Daniela Spampinato, Concetto Pennisi, Matteo |
| author_facet | Russo, Chiara Maria Carnemolla, Simone Palazzo, Simone Giordano, Daniela Spampinato, Concetto Pennisi, Matteo |
| contents | Interpreting the decisions of deep image classifiers remains challenging, particularly in black-box settings where model internals are inaccessible. We introduce OCCAM, a framework for open-set causal concept explanation and ontology induction in vision models. OCCAM discovers visual concepts in an open-set manner, localizes them via text-guided segmentation, and performs object-level interventions by removing concepts to measure changes in class confidence, estimating each concept's causal contribution. Beyond local explanations, OCCAM aggregates interventional evidence across a dataset to induce a structured concept ontology that captures how classifiers globally organize visual concepts. Reasoning over this ontology reveals consistent dependencies between concepts, exposes latent causal relations, and uncovers systematic model biases. Experiments on Broden and ImageNet-S across multiple classifiers show that OCCAM improves explanation quality in open-set black-box settings while providing richer global insight than per-image attribution methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_18481 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | OCCAM: Open-set Causal Concept explAnation and Ontology induction for black-box vision Models Russo, Chiara Maria Carnemolla, Simone Palazzo, Simone Giordano, Daniela Spampinato, Concetto Pennisi, Matteo Artificial Intelligence Interpreting the decisions of deep image classifiers remains challenging, particularly in black-box settings where model internals are inaccessible. We introduce OCCAM, a framework for open-set causal concept explanation and ontology induction in vision models. OCCAM discovers visual concepts in an open-set manner, localizes them via text-guided segmentation, and performs object-level interventions by removing concepts to measure changes in class confidence, estimating each concept's causal contribution. Beyond local explanations, OCCAM aggregates interventional evidence across a dataset to induce a structured concept ontology that captures how classifiers globally organize visual concepts. Reasoning over this ontology reveals consistent dependencies between concepts, exposes latent causal relations, and uncovers systematic model biases. Experiments on Broden and ImageNet-S across multiple classifiers show that OCCAM improves explanation quality in open-set black-box settings while providing richer global insight than per-image attribution methods. |
| title | OCCAM: Open-set Causal Concept explAnation and Ontology induction for black-box vision Models |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2605.18481 |